Improved attention mechanism and residual network for remote sensing image scene classification

Jiayuan Kong, Yurong Gao, Yanjun Zhang, Huimin Lei, Yao Wang, Hesheng Zhang

Research output: Contribution to journalArticlepeer-review


Aiming at the problem that the accuracy of traditional remote sensing image classification model is not ideal, a classification method based on improved attention mechanism and residual network is proposed. In order to prevent overfitting, we choose the network structure of ResNet18 as the framework. Meanwhile, ResNet18 has a small number of parameters and a fast calculation speed. Then, a parallel attention mechanism, ProCBAM, is proposed and added to the BasicBlock of the residual network. By optimizing the representation of the feature map from the spatial dimension and channel dimension of the feature map, more detailed image information is learned and image recognition errors are reduced. The experimental results on the open source dataset show that the accuracy is 93.86%, and a more accurate image classification model is successfully trained.

Original languageEnglish (US)
Pages (from-to)134800-134808
Number of pages9
JournalIEEE Access
StatePublished - 2021


  • Attention mechanism
  • Image classification
  • Remote sensing
  • Residual network
  • Spatial information

ASJC Scopus subject areas

  • General Computer Science
  • General Materials Science
  • General Engineering
  • Electrical and Electronic Engineering


Dive into the research topics of 'Improved attention mechanism and residual network for remote sensing image scene classification'. Together they form a unique fingerprint.

Cite this